Inference
EN
JetBrains

Mellum2-12B-A2.5B-Instruct

Mellum2-12B-A2.5B-Instruct is a standard model from JetBrains in the Mellum2 12B-A2.5B family. It has 6 scored results in 5 categories. It has no index.

  • Open weights
  • Standard
  • Input: text
  • Released 2026-05
UnrankedEnough results for an index.
IndexAll categories, out of 100.UnrankedMedian 51.0
CoverageIndex weight that has results.65%
SpeedOutput tokens per second.—Median 67/s
First tokenSeconds to the first output token.—Median 1.3 s
Input / 1MUS dollars per 1M input tokens.—Median $0.60
Output / 1MUS dollars per 1M output tokens.—Median $2.55
ContextMaximum tokens in one request.128KMedian 262K
EloLMArena rating and rank.—Median 1422

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.1 counted26.2—
CodingCode writing and repair.1 counted15.0—
ReasoningLogic problems and puzzles.None counted——
MultimodalTasks with images and text.None counted——
KnowledgeFacts and expert knowledge.2 counted18.7—
MultilingualTasks in many languages.1 counted19.1—
InstructionTasks with strict prompt rules.1 counted15.8—
MathMath problems.None counted——

Results

6 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
EuroEval GermanMultilingual31.9% ±0.7115 / 191
EuroEval DutchMultilingual37.1% ±0.6108 / 175
Berkeley Function Calling Leaderboard v4Agentic44.2%21 / 26
Instruction-Following EvalInstruction75.8%62 / 70
Graduate-Level Google-Proof Q&AKnowledge40.9%132 / 141
MMLU-ReduxKnowledge78.1%18 / 19
LiveCodeBench v6Coding37.2%56 / 59
GPQA Diamond · BenchLMKnowledge40.9%82 / 84

Sources

BenchLMCC BY-NC 4.0EuroEvalMIT

Questions

What is the context window of Mellum2-12B-A2.5B-Instruct?
128K tokens.
Is Mellum2-12B-A2.5B-Instruct open source?
Yes. Its weights are published.
When was Mellum2-12B-A2.5B-Instruct released?
May 2026.

More from JetBrains

All models
Mellum2-12B-A2.5B-ThinkingUnrankedMellum2.1-12B-A2.5B-ThinkingUnranked